Exploratory Data Analysis and Data Visualisation

policing_data <- read.csv("./37-00049_UOF-P_2016_prepped.csv") # all the visualizations drawn from this file

Problem Statement?

How do you measure justice? And how do you solve the problem of racism in policing? We look for factors that drive racial disparities in policing by analyzing census and police department deployment data. The ultimate goal is to inform police agencies where they can make improvements by identifying deployment areas where racial disparities exist and are not explainable by crime rates and poverty levels.

Summary about the analyses and visualizations

This project shows different visualizations and analysis of the dataset about policing equity in 2016 through different factors. This dataset has 2383 rows and 47 columns where different results have been concluded racially and it has been checked that how different crimes have happened throughout the year through which police department can learn and grow accordingly. All of the key findings from the project are defined below where I have concluded different entities from the dataset with their properties or characteristics:

Entities Characteristics/Features
Incident (date, time, reason)
Officer (ID, Gender, Race, Hire_Date, Years_on_Force, Injury, Injury_Type, Hospitalization)
Subject (ID, Race, Gender, Injury, Injury_Type, Was_Arrested, Description, Offense)
Street (Number, Name, Directions, Type)
Location (Reporting_Area, Sactor, Division, District, Full_street_address, City, State, Latitude, Longitutde,)
Force (Reason, type1, type2, type3, type4, type5, type6, type7, type8, type9, type10, effectiveness, EC_Cycles, UOF_number(use_of_force_number),Beat)

These entities will be used to assess different visualization to make more safety in public and to inform law makers that how they are acting about different equities. Using these entities I have made different relations and have concluded different analysis and conclusions from them which are different visualizations of level 7 interactively, or animated graphs shown below. Now on compare we can found out how these entities are affecting each other and how we can utilize their information to find out un-justice or racism.

Racial and Gender Proportions

Checking officer and subject gender and race from which clearly have concluded that male are covering maximum of the proportion from both sides which is around 80% of the population. Other officers are White, Hispanic and Black covering 61%, 20%, 14% of the whole population accordingly whereas subject races are Black, Hispanic, and White covering the 55%, 21% and 19% of the whole population accordingly.

Different Actions Against Subject

 
                Arrest        Call for Cover     Crime in Progress 
                  48.6                   5.5                   3.4 
         Crowd Control   Off-Duty Employment     Off-Duty Incident 
                   0.2                   2.1                   0.4 
                 Other Other ( In Narrative)       Pedestrian Stop 
                   0.5                   2.9                   1.5 
          Service Call   Suspicious Activity          Traffic Stop 
                  28.4                   2.0                   3.9 
     Warrant Execution 
                   0.4

These analyses shows that most of the actions which taken place were arrest, service call, Call for cover, traffic stop, and crime in progress were covering around 48%, 28.4%, 5.5%, 3.9% and 3.4% of the total crimes where different races got affected by it. Black subjects got arrested, service call, traffic stop and call for cover a lot. Black, Hispanic and white arrested around 673, 244 and 215 of the total crimes. Then so on service call was distributed around 360, 161 and 140 between Black, White and Hispanic.

Officers Races Caught Subject Races

 
1 Asian                     2.3
2 Black                    14.1
3 Hispanic                 20.3
4 Other                     1.5
5 White                    61.7

After seeing these analysis I checked that how different races of Officers are impacting other races of Subjects. In 2016 61.7% Black, Hispanic and White people get caught by White people. Out of this percentage, White caught around 673 Black, 302 Hispanic and 277 White subjects. Hispanics and Black caught people with the percentage of 20.3% and 14.1% where Hispanic caught 230 Black and Black officers caught 201 Black subjects.

Crimes through out the year over Calendar

optional caption text Plotting the calendar for the year to see that how many crimes has happened through out the year. Where it was clearly seen that on certain dates Dec 14 2016, Jun 06 2016, May 01 2016, May 27 2016 and Aug 24 2016 when the crime rates were touching 20 or 20+. December, and April was the months where it could seen that crimes are around zero’s on most of the days. * This plot is more visible if you inspect it in new window.

Yearly Crimes Rate through line plot

Drawn two differnet graphs to check how over the year incidents are decreasing. Where it was seen that the maximum crimes in a day happened in October 2016 where it even crosses 20 crimes and total of 24 crimes happened. On different occasions different number of crimes happened but overall crimes ration was below average and it was seen that overall till the end of 2016 crimes were decreasing with the significant numbers.

Incidents Happened in Different Divisions

Incidents MAP of Divisions (Long, Lat)

Dallas, Texas Map with crimes

Crimes happened in the whole dellas was shown out through the map to see which area got affected most. It was seen that Central area of Dellas was on top with the most crime of 563 and then SouthEast area was second highest with most crimes of 362. And it was clearly seen that overall crimes are decreasing in all the divisions but 2 divions are still constant where Central and NorthEast was on top. *This graph is viewable from the viewer section if we run it through console.

Officers Got Injured and Bruises

It was clearly seen that not most of the officers got injured during different crimes but 234 got injured about where they got mostly two types of scrapes from Abrasion/Scrape to No Visible Injury. In case of Subjects those got injured more 629 totally and and they also got Abrasion/Scrape, Laceration/Cut and Puncture around 189, 52, 24 respectively out of the totall subjects caught.

Officers on force from their joining

Subjects Offenses they have made

Checking the top 10 most happening offenses by the subjects where APOWW, Publix Intoxication, Warrant/Hold or Assault/FV was the most happening offenses with 399, 179, 110, and 92 numbers recorded accordingly. In the data it is shown that there is total of 6 types of forces used where mostly just upto force type 3 were used but then it was stopped.

Conclusion:

It is concluded through the data that mostly crimes were happened by the subjects Black and Hispanic. Mostly White officers were in actions and caught those Black subjects. Officers of type Black and White were on top in their force years tenure. Overall crimes were decreasing throughout the year till the end of 2016 and most of the crimes has happened in the Central area of the Dellas. Subjects were got injured more than officers where they committed APOWW and Intoxication offense alot while Officers got Abrasion/Scrape alot.

References: